The Devil Is in the Verb

Superforecaster Ryan Adler on what Polymarket’s Venezuela kerfuffle reveals about “simple” forecasting questions.

Writing questions is harder than it looks. For simple things, like run-of-the-mill economic data (when the government isn’t shut down), it is rather easy. However, we do not live in a simple world, and framing forecasting questions for highly relevant things can require extremely methodical work. Having written well north of 4,000 forecasting questions and forecast on hundreds more, I’ve suffered through every mistake imaginable (and probably more to come).

After President Trump arranged for Nicolas Maduro to receive free housing in New York (Mayor Mamdani should be proud), Polymarket found itself in a bit of a pickle of its own making.

Will the U.S. invade Venezuela by [date]?

It’s a simple question, a short sentence, and something that has been on a whole lot of minds in recent months.

Straightforward enough? At first glance, it might seem black-and-white. Folks at Polymarket apparently thought so, but they were quite mistaken. Many market participants assumed that the US operation constituted an invasion, and Polymarket has concluded otherwise.

As is often the case for a forecasting question, which is functionally identical to an event contract, the devil is in the operative verb: invade.

In what we at Good Judgment would call the resolution criteria, Polymarket attempted to expound a bit on what would constitute the US invading Venezuela:

This market will resolve to “Yes” if the United States commences a military offensive intended to establish control over any portion of Venezuela between November 3, 2025, and January 31, 2026, 11:59 PM ET. Otherwise, this market will resolve to “No”.

Left only with this language, I have to say the traders who think they should get paid out based on the events surrounding Operation Absolute Resolve have a very strong case to make. Was this a military offensive? Dozens of dead Cuban intelligence officers would say so, and I don’t think anyone would claim otherwise. Did that offensive “intend to establish control over any portion of Venezuela” (emphasis added)? Absolutely. Maduro was in a building, which is on land. According to all reporting, US forces landed on Venezuelan territory and clearly prevented Maduro from escaping the structure he and his wife were in. A perimeter was established. What’s in that perimeter? A portion of Venezuela.

Sure, this would have been a very small chunk of Venezuela, and the control established (intention for doing so is implicit) was of a very short duration. However, Polymarket didn’t take the time to add modifiers that would have excluded the operation that we saw.

As an undocumented attorney (got my JD back in the day, but never practiced), one of my first classes was on contracts. While there are many nuances in interpretation, enforceability, and public policy from jurisdiction to jurisdiction, there is one thing that controls a contract: its text. Here, Polymarket may have intended for its language to mean something different than the seizure of Maduro, but that’s not what they wrote in the contract they offered to the market.

Mistakes will be made, but this, in my opinion and experience, was an unforced error.

* Ryan Adler is a Superforecaster, GJ managing director, and leader of Good Judgment’s question team

Don’t Take the Bait

Superforecaster Ryan Adler shares a childhood memory to discuss what bottom-feeders can teach us about consuming social media.

I’m sure everyone has memories of unremarkable childhood moments that remain vivid decades later. When I was a wee one in the ’80s, probably Reagan’s second term, I recall to this day a funny experience I had on a marina back in Jackson County, Missouri. Lake Jacomo (as in Jackson County, MO…not the most creative name) is a pretty big lake with all the usual finned critters in any Midwestern body of water: bluegill, bass, crappie, catfish, and, of course, carp.

As living garbage disposals, the carp will congregate where they expect to find anything to eat. Marinas are a great place for that, since they are not picky. People throw out unused bait, scraps from cleaning the day’s catch, and plenty of other things into the water. In this particular spot, if anything touched the surface, fat carp would swim on top of each other to get a bite. When I was there one day, I saw a used cigarette butt on the dock. So, I did a little experiment. I threw the butt into the water where I had seen the carp congregating. The fifteen-second battle that ensued has stayed with me ever since. These creatures lived for trash, any trash, that might come their way.

One Carp, Two Carp, Red Carp, Blue Carp

This carp-like behavior isn’t new among humans, but the advent of social media has enabled anyone to throw cigarette butts at millions of people with zero regard for factual accuracy or good-faith characterizations of events. Vast swaths of X, Facebook, Bluesky, and others have become marina feeding grounds for intellectually lazy people who worship at the altar of induction. Sure, some sites have more blue carp than red carp and every political color of carp can find the trash it seeks, but you still end up with a product that will leave a nasty taste in your mouth.

This phenomenon has wide-ranging implications for a lot of things, but anyone who wants to be taken seriously as a forecaster needs to be actively aware of it. Social media is filled with narrative fallacy bait cut by people with an agenda, aiming solely to make people believe not as the world is, but as they want it to be (or, in many cases, simply lying for the sake of lying). Do you want to be a good forecaster? Don’t be a carp. Learn to be picky about what you bite!

* Ryan Adler is a Superforecaster, GJ managing director, and leader of Good Judgment’s question team